agreed that comparing just on class type does seem very
fragile/specific.
/max
Hello,
I'm writing a JUnit test to verify that a quickfix for a JAX-RS
problem works fine, but my test fails because the change is not
applied during the call to Change#perform().
When digging into the code, I found that my problem is in the
following method of
org.jboss.tools.common.refactoring.MarkerResolutionUtils:
@SuppressWarnings("unchecked")
public static <T extends IJavaElement> T
findWorkingCopy(ICompilationUnit compilationUnit, T element) throws
JavaModelException{
if(element instanceof IAnnotation){
IJavaElement parent = findWorkingCopy(compilationUnit,
element.getParent());
if(parent instanceof IAnnotatable){
for(IAnnotation a : ((IAnnotatable)parent).getAnnotations()){
if(a.getElementName().equals(element.getElementName()))
return (T)a;
}
}
}else if(element instanceof ILocalVariable && ((ILocalVariable)
element).isParameter()){
IJavaElement parent = findWorkingCopy(compilationUnit,
element.getParent());
if(parent instanceof IMethod){
for(ILocalVariable parameter : ((IMethod)parent).getParameters()){
if(parameter.getElementName().equals(element.getElementName()) &&
parameter.getTypeSignature().equals(((ILocalVariable)element).getTypeSignature()))
return (T)parameter;
}
}
}else{
IJavaElement[] elements = compilationUnit.findElements(element);
if(elements != null){
for(IJavaElement e : elements){
if(e.getClass().equals(element.getClass()))
return (T)e;
}
}
}
return null;
}
In the code above, the last if/else block checks if
e.getClass().equals(element.getClass()), which does not work in my
case because I have org.eclipse.jdt.internal.core.SourceType vs
org.eclipse.jdt.internal.core.ResolvedSourceType. I wonder if we
should not replace this part with a comparison of a comparision of
getHandleIdentifier() of both elements, which would make more sense, I
believe.
For example, in my unit tests, it would be:
=org.jboss.tools.ws.jaxrs.tests.sampleproject/src\/main\/java<org.jboss.tools.ws.jaxrs.sample.services{BAZ.java[BAZ
in both cases ;-)
WDYT ?
If the change I suggest makes sense, I'll be more than happy to open a
JIRA and attach a PR to it ;-)
Thanks.
Best regards,
/Xavier
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev